Max-Margin Synchronous Grammar Induction for Machine Translation

Traditional synchronous grammar induction estimates parameters by maximizing likelihood, which only has a loose relation to translation quality. Alternatively, we propose a max-margin estimation approach to discriminatively inducing synchronous grammars for machine translation, which directly optimizes translation quality measured by BLEU. In the max-margin estimation of parameters, we only need to calculate Viterbi translations. This further facilitates the incorporation of various non-local features that are defined on the target side. We test the effectiveness of our max-margin estimation framework on a competitive hierarchical phrase-based system. Experiments show that our max-margin method significantly outperforms the traditional twostep pipeline for synchronous rule extraction by 1.3 BLEU points and is also better than previous max-likelihood estimation method.

• Publication year

  2013

• Venue

  Conference on Empirical Methods in Natural Language Processing

• Publication date

  2013-10-01

• Fields of study

  Computer Science

• Identifiers
  DOI 10.18653/v1/d13-1026
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Maximum Expected BLEU Training of Phrase and Lexicon Translation Models

  2012cited by this paper
• Unsupervised Discriminative Induction of Synchronous Grammar for Machine Translation

  2012influential reference
• A Bayesian Model for Learning SCFGs with Discontiguous Rules

  2012cited by this paper
• Structured Ramp Loss Minimization for Machine Translation

  2012cited by this paper
• An Unsupervised Model for Joint Phrase Alignment and Extraction

  2011cited by this paper
• Feature-Rich Language-Independent Syntax-Based Alignment for Statistical Machine Translation

  2011cited by this paper
• Fast Generation of Translation Forest for Large-Scale SMT Discriminative Training

  2011cited by this paper
• The CMU-ARK German-English Translation System

  2011cited by this paper
• Joint Parsing and Alignment with Weakly Synchronized Grammars

  2010cited by this paper
• Feature-Rich Discriminative Phrase Rescoring for SMT

  2010cited by this paper
• Unsupervised Syntactic Alignment with Inversion Transduction Grammars

  2010cited by this paper
• Discriminative Modeling of Extraction Sets for Machine Translation

  2010cited by this paper
• Learning Translation Boundaries for Phrase-Based Decoding

  2010cited by this paper
• A Bayesian Model of Syntax-Directed Tree to String Grammar Induction

  2009cited by this paper
• Variational Decoding for Statistical Machine Translation

  2009cited by this paper
• Sinuhe – Statistical Machine Translation using a Globally Trained Conditional Exponential Family Translation Model

  2009cited by this paper
• A Gibbs Sampler for Phrasal Synchronous Grammar Induction

  2009cited by this paper
• Effective Use of Linguistic and Contextual Information for Statistical Machine Translation

  2009cited by this paper
• 11,001 New Features for Statistical Machine Translation

  2009cited by this paper
• Sampling Alignment Structure under a Bayesian Translation Model

  2008cited by this paper
• Bayesian Learning of Non-Compositional Phrases with Synchronous Parsing

  2008cited by this paper
• Forest Reranking: Discriminative Parsing with Non-Local Features

  2008influential reference
• A Discriminative Latent Variable Model for Statistical Machine Translation

  2008cited by this paper
• Syntactic Re-Alignment Models for Machine Translation

  2007cited by this paper
• Inversion Transduction Grammar for Joint Phrasal Translation Modeling

  2007cited by this paper
• Moses: Open Source Toolkit for Statistical Machine Translation

  2007influential reference
• Online Large-Margin Training for Statistical Machine Translation

  2007cited by this paper
• Hierarchical Phrase-Based Translation

  2007influential reference
• Discriminative Word Alignment with Conditional Random Fields

  2006cited by this paper
• ORANGE: a Method for Evaluating Automatic Evaluation Metrics for Machine Translation

  2004influential reference
• Max-Margin Parsing

  2004cited by this paper
• Statistical Significance Tests for Machine Translation Evaluation

  2004cited by this paper
• Improvements in Phrase-Based Statistical Machine Translation

  2004cited by this paper
• Minimum Error Rate Training in Statistical Machine Translation

  2003influential reference
• Statistical Phrase-Based Translation

  2003cited by this paper
• SRILM - an extensible language modeling toolkit

  2002cited by this paper
• A Phrase-Based,Joint Probability Model for Statistical Machine Translation

  2002cited by this paper
• Bleu: a Method for Automatic Evaluation of Machine Translation

  2002cited by this paper
• Mathematical Programming manuscript No. (will be inserted by the editor) Pegasos: Primal Estimated sub-GrAdient SOlver for SVM

  year unknowncited by this paper

• Optimization for Statistical Machine Translation: A Survey

  2016cites this paper
• A pilot study towards end-to-end MT training

  2015cites this paper
• Synchronous Constituent Context Model for Inducing Bilingual Synchronous Structures

  2014cites this paper
• Topic-Based Dissimilarity and Sensitivity Models for Translation Rule Selection

  2014cites this paper
• Machine Translation: Future Directions

  2013cites this paper